Coffee is good for you. Specialty, more so.

Across dozens of large prospective cohorts and umbrella reviews, drinking three to four cups of coffee a day is associated with a lower risk of all-cause mortality, cardiovascular disease, type 2 diabetes, Parkinson's disease, Alzheimer's disease, liver cirrhosis, liver cancer and several other cancers, when compared with not drinking coffee at all.

The 2017 BMJ umbrella review by Poole and colleagues — still the most rigorous synthesis to date — concluded that coffee consumption at three to four cups a day is 'more likely to benefit health than to harm it' across almost every outcome studied.

None of this is medical advice. But the honest reading of the science is that coffee, for most adults, is a net positive — not a guilty pleasure to be justified.

Most published health research uses 'coffee' as a single category, which is generous to commodity coffee and unfair to specialty. In practice, a cup of freshly roasted, traceable specialty coffee and a cup of stale, over-roasted commodity blend share a name and very little else.

Specialty coffee is graded 80+ on the Specialty Coffee Association scale — meaning it is free of the primary defects (black beans, sour beans, insect damage, mould) that are common in commercial-grade lots. Those defects are not only flavour problems. Mouldy or improperly stored coffee is the primary source of ochratoxin A, a mycotoxin that is essentially the reason coffee has any 'toxin' reputation at all. In sorted, traceable specialty lots, OTA is typically at or below the limit of detection.

Specialty coffee is also fresh. Antioxidants and chlorogenic acids — the compounds most consistently linked to coffee's benefits — degrade with time and with heat. A bag roasted last week and brewed today carries far more of the bioactive fraction than a supermarket blend roasted last quarter and stale on the shelf.

Roast level is not a matter of taste alone. It changes the chemistry of the cup.

Chlorogenic acids, the largest family of polyphenols in coffee and the compounds most often invoked for coffee's anti-inflammatory and metabolic effects, are heat-sensitive. Light roasts retain a substantially higher fraction than dark roasts — measured losses of 60–90% at the darker end of the scale are typical in the peer-reviewed literature (Farah, Moreira and others).

Acrylamide, a compound of regulatory interest, actually peaks early in the roast and then declines — so a very light roast is not automatically 'cleaner' on that axis — but the polycyclic aromatic hydrocarbons and other pyrolysis products that accumulate as beans are pushed into second crack and beyond are clearly higher in dark roasts. On balance, a well-developed light roast preserves more of what the science says is beneficial and generates less of what it says is not.

Flying Frog roasts every coffee light for exactly this reason. It is the profile that lets the origin speak, and it is also the profile the research quietly rewards.

Coffee causes cancer. Withdrawn: the WHO's IARC removed coffee from its 'possibly carcinogenic' list in 2016 after reviewing more than a thousand studies, and reclassified it as unlikely to be carcinogenic. The signal that put it on the list originally came from very hot beverages in general, not from coffee itself.

Coffee is bad for your heart. Not in moderate amounts. Meta-analyses consistently show a J-shaped or inverse relationship between coffee intake and cardiovascular mortality, with the lowest risk around three to five cups per day. Unfiltered brewing (French press, cezve) can raise LDL cholesterol via diterpenes; paper-filtered brewing removes them almost entirely.

Coffee dehydrates you. It doesn't, at habitual doses. The mild diuretic effect of caffeine is more than offset by the water in the cup. Regular drinkers show no net dehydration in controlled studies.

Coffee stunts your growth. There is no credible evidence for this claim in humans, and never was.

Decaf is unhealthy. It is not — when it is made with modern water or naturally derived processes. The safe methods are Swiss Water, Mountain Water, and sugarcane ethyl acetate: they leave essentially no solvent residue and retain most of the polyphenol content. Cohort studies associate decaf with many of the same metabolic and cardiovascular benefits as regular coffee. The method to avoid is methylene chloride (dichloromethane), a synthetic solvent still used in some commodity decaf but essentially gone from specialty coffee.

Multiple large prospective cohorts have observed a lower incidence of dementia and Alzheimer's disease among adults who drink coffee regularly compared with non-drinkers. A pooled analysis in the UK Biobank cohort (over 300,000 adults, followed for more than a decade) found the lowest risk of dementia among those consuming roughly two to three cups per day.

Proposed mechanisms are plausible but not settled: chlorogenic acids and other polyphenols with antioxidant activity in the brain; adenosine-receptor modulation by caffeine; a modest reduction in the accumulation of beta-amyloid in animal models.

The honest reading: at moderate intake, coffee is associated with a lower risk of cognitive decline, and there is no credible signal that it accelerates it.

The strongest single-outcome signal in coffee epidemiology is type 2 diabetes. A dose-response meta-analysis by Ding and colleagues (Diabetes Care, 2014) reported a roughly 9% lower risk of type 2 diabetes per additional cup per day, holding across caffeinated and decaffeinated coffee — implicating the non-caffeine fraction (chlorogenic acids, trigonelline) as at least partly responsible.

Coffee drinkers also show consistently lower rates of non-alcoholic fatty liver disease, liver fibrosis and hepatocellular carcinoma. The European Association for the Study of the Liver's 2016 clinical guidance explicitly recommends coffee for patients with chronic liver disease.

Prospective studies — most notably the Health Professionals Follow-Up Study (~50,000 men) and several European cohorts pooled by the International Agency for Research on Cancer — have repeatedly reported an inverse association between coffee consumption and the risk of aggressive or lethal prostate cancer.

A 2021 meta-analysis in BMJ Open (16 cohort studies, over one million men) reported a 9% lower risk of prostate cancer per additional cup per day, with a stronger inverse association for advanced disease.

Buy fresh. A bag roasted within the last few weeks carries more of the bioactive fraction than one that has sat on a shelf for months.

Choose light. A well-developed light roast preserves chlorogenic acids and produces less of the compounds generated at very dark roast levels.

Filter through paper. It removes the diterpenes (kahweol, cafestol) that raise LDL cholesterol without removing the polyphenols that benefit it.

Grind before you brew. Ground coffee stales within minutes.

Two to four cups a day is where population-level benefits are most consistent and side effects least common. A 'cup' in research is typically 150 ml — closer to a small filter than a takeaway cup.

Pregnant people, people with certain arrhythmias, and people on specific medications should follow clinical guidance rather than population averages.

Poole R. et al. Coffee consumption and health: umbrella review of meta-analyses. BMJ, 2017. Ding M. et al. Caffeinated and decaffeinated coffee consumption and risk of type 2 diabetes: a systematic review and dose-response meta-analysis. Diabetes Care, 2014. Farah A. and Donangelo C.M. Phenolic compounds in coffee. Brazilian Journal of Plant Physiology, 2006. Loomis D. et al. Carcinogenicity of drinking coffee, mate, and very hot beverages. IARC / Lancet Oncology, 2016. Cornelis M.C. et al. Genetics of coffee, cognition and dementia. UK Biobank analyses, 2020–2023. Wilson K.M. et al. Coffee consumption and prostate cancer risk and progression, Health Professionals Follow-Up Study. Chen X. et al. Coffee consumption and risk of prostate cancer: a meta-analysis of prospective studies. BMJ Open, 2021. EASL Clinical Practice Guidelines on non-alcoholic fatty liver disease, 2016.