An honest, citation-driven walk-through of what coffee does (and doesn't) do to your body. Parkinson's, liver, type 2 diabetes, heart failure, cancer, dementia, longevity — and where to skip the cup. 9 minutes.
For most adults, the dose-response curves on moderate coffee consumption (3–4 cups a day) bend in the direction you'd want: lower risk of Parkinson's, type 2 diabetes, liver disease, certain cancers, and all-cause mortality. The effect sizes for some of these (Parkinson's especially) are larger than most pharmacological interventions can claim.
The cases where you should drink less or skip coffee — pregnancy, severe insomnia, anxiety disorders, certain medications — are real and worth taking seriously, but they're narrower than the cautious advice you'll often read makes them sound.
Almost everything we know about coffee and disease risk comes from prospective cohort studies: large groups of people followed over decades, who report their habits, and whose outcomes are tracked. This is observational evidence. It can tell you that coffee drinkers get less Parkinson's. It can't, by itself, prove the coffee is what's protecting them — they might be different in other ways too.
What gives confidence here is the weight of replication. The coffee-Parkinson's finding holds up across European, Asian, and US cohorts, and the proposed biological mechanism (caffeine antagonizing adenosine A2A receptors on dopaminergic neurons) has been confirmed in animal models and small human trials. When a finding replicates across continents and has a plausible mechanism, the observational data starts to mean something close to causation.
This is the most documented health benefit of coffee, and it's why 1% of every ShopCoffee bag goes to the Michael J. Fox Foundation for Parkinson's Research.
Across multiple prospective cohorts, people who drink 3–5 cups of coffee a day are 30–60% less likely to develop Parkinson's disease than non-drinkers. The Harvard School of Public Health's analyses have repeatedly landed in this range; NIH-funded replications agree. The effect is dose-responsive (more coffee, more protection, up to about 5 cups a day), and it's larger in men than in women (where post-menopausal hormone therapy appears to interact with the caffeine pathway).
The mechanism is well-characterized. Caffeine is an adenosine receptor antagonist. The A2A subtype of adenosine receptors sits on dopaminergic neurons in the basal ganglia — exactly the neurons that die off in Parkinson's. Caffeine occupies those receptors, modulates dopamine signaling, and appears to be neuroprotective. The animal-model evidence for this is consistent and the human pharmacology lines up.
It is not a cure. It does not protect everyone. It is one of the most robust risk-reduction findings in dietary epidemiology, and the foundation we donate to is the largest nonprofit funder of Parkinson's drug development in the world.
Coffee drinkers have lower rates of nonalcoholic fatty liver disease (NAFLD), liver fibrosis, cirrhosis, and hepatocellular carcinoma (liver cancer). This holds even in heavy drinkers, even in chronic hepatitis C patients, and the protective association generally strengthens with dose up to about 4 cups a day.
The mechanism is less neat than Parkinson's — likely a combination of caffeine plus several non-caffeine compounds (chlorogenic acids, diterpenes) reducing liver inflammation and modulating fat metabolism. Decaf coffee also shows protective effects against liver disease, which suggests the benefit isn't purely caffeine-driven.
The American Association for the Study of Liver Diseases has gone on the record that coffee consumption is associated with reduced liver disease risk. The American Liver Foundation echoes this.
Each additional cup of coffee per day is associated with about a 6–7% reduction in risk of developing type 2 diabetes. That's a small per-cup number, but it compounds: someone drinking 4 cups a day, every day for 20 years, is meaningfully less likely to develop T2D than someone drinking none.
The mechanism appears to involve improved insulin sensitivity and lower systemic inflammation, plus possibly direct effects of chlorogenic acid on glucose absorption. Decaf again shows the same direction of effect, again suggesting it's not just caffeine.
This one used to be controversial. The current evidence is mostly reassuring. The American Heart Association published an explicit "Coffee lovers, rejoice" piece in 2021 noting that each additional daily cup of coffee was associated with a 5–12% reduction in heart failure risk across three large cohorts, with the protective effect persisting up to about 4 cups a day.
The arrhythmia story is also clearer than it used to be: moderate coffee consumption does not increase the risk of atrial fibrillation, and may slightly reduce it. Older worries about coffee and high blood pressure mostly evaporate in long-term studies — there's a small acute pressor effect from caffeine, but no chronic blood pressure increase in regular drinkers.
The exception is if you have a known cardiac arrhythmia that's caffeine-sensitive (some people genuinely have these), in which case listen to your cardiologist, not us.
In 2016, the World Health Organization's International Agency for Research on Cancer removed coffee from its list of "possibly carcinogenic" substances after a thorough review. The current state of evidence is that moderate coffee consumption is associated with reduced risk of several cancers: liver, endometrial, colorectal, and possibly oral and pharyngeal.
The one nuance: very hot beverages (hotter than ~149°F / 65°C) of any kind are classified as probably carcinogenic to the esophagus. This isn't a coffee problem — it's a "drinking scalding-hot liquids" problem — but if you swallow your coffee straight off the kettle, let it cool to drinking temperature.
The evidence here is suggestive but less rock-solid than Parkinson's. Multiple cohort studies have found that regular coffee drinkers have lower rates of Alzheimer's disease and general dementia in late life, with effect sizes in the 20–30% range at moderate consumption. The mechanism is hypothesized to overlap with the Parkinson's pathway (adenosine receptor modulation in the brain) plus possible reduction in tau and amyloid accumulation in animal models.
The Alzheimer's Association acknowledges the association but rightly notes it doesn't yet rise to a clinical recommendation. Worth knowing about; not worth treating coffee as a medication.
Pulled back to the highest level, large prospective studies (NIH-AARP, UK Biobank, several European cohorts) consistently find that moderate coffee drinkers have lower all-cause mortality than non-drinkers. The hazard ratio runs roughly 0.85–0.92 across these studies — that is, about 8–15% lower risk of dying in any given year, after controlling for smoking, obesity, alcohol, and the obvious confounders.
Decaf shows similar (slightly smaller) longevity associations. That's important because it suggests something in coffee beyond caffeine — the polyphenols, the chlorogenic acids — is doing real work.
Pregnancy. ACOG and most obstetric guidelines recommend keeping caffeine under about 200mg a day during pregnancy. That's one strong cup. The evidence on harm at moderate intake is mixed, but it's a defensible "when in doubt, less" call.
Severe insomnia or sleep-onset disorders. Caffeine has a half-life of 5–7 hours in most adults. If you genuinely struggle to fall asleep, cut off coffee by early afternoon. The "I can drink espresso after dinner and sleep fine" thing is real for some people and a lie they tell themselves for others.
Diagnosed anxiety disorders. Caffeine can amplify panic and generalized anxiety symptoms in people predisposed to them. If you've noticed your morning cup makes you jittery beyond ordinary alertness, listen to that.
Certain medications. Caffeine interacts with several prescription drugs, including some antibiotics (ciprofloxacin), some antidepressants (MAOIs especially), some asthma medications (theophylline), some thyroid medications, and — relevant for some of our readers — blood-thinning anticoagulants like warfarin. If you're on a chronic medication, ask your pharmacist about caffeine interactions. The answer is usually fine, occasionally not.
People with antiphospholipid syndrome (APS) or on long-term anticoagulation. Caffeine doesn't directly worsen APS, but it can interact with the anticoagulants APS patients typically take. Not a hard contraindication — a "talk to your hematologist about your specific regimen" item.
Unfiltered coffee — French press, Turkish, espresso to a lesser degree — contains two diterpenes (cafestol and kahweol) that paper filtering removes. In large daily quantities, these compounds can modestly raise LDL cholesterol. For most people drinking 1–3 cups a day this is negligible. For someone drinking 6+ cups of unfiltered French press a day with already-high cholesterol, it's worth knowing.
The cure is simple: a paper filter. Drip and pour-over remove almost all the diterpenes. If you mainly drink French press or espresso and you've got cholesterol issues, consider switching some of your daily volume to filtered brewing.
None of this is medical advice. If you're managing a condition or taking medication, your clinician's answer trumps any internet article — including this one.