Five cups of coffee are safe. But not for everyone.

What the AHA Scientific Statement from July 2026 actually says. And why the new number is only the beginning of a meaningful recommendation.

What the statement actually says

On 20 July 2026, the American Heart Association published in Circulation the most comprehensive position paper on caffeine and cardiovascular health in years. Not a single finding, but a structured evaluation of the entire available evidence — developed by an author team led by Gregory Marcus (UCSF) and Frank Hu (Harvard).

The result is clearer than many would have expected: up to 400 mg of caffeine per day — roughly equivalent to three to five cups of black coffee without additives — is not only safe for most adults, but frequently associated with measurable benefits.

Moderate coffee consumption converges across large cohort studies and smaller randomised interventions with a lower risk for several cardiovascular conditions:

  • Coronary heart disease
  • Stroke
  • Heart failure
  • Atrial fibrillation
  • High blood pressure — from three cups per day
  • Type 2 diabetes

Effect sizes vary by endpoint. What the authors explicitly note: there is no one-size-fits-all recommendation. Individual responses to caffeine are too varied to apply a single number to everyone.

That is the most important sentence in the entire statement. And it tends to get lost in the headline noise.

Where the nuance lies

Three points that are relevant in practice. They rarely appear in coverage.

Energy drinks are not coffee. Per unit of volume they contain three to four times more caffeine than filter coffee, along with sugar and additional stimulants. The AHA authors explicitly warn of cardiac arrhythmias — documented even in healthy adults under 30. Practitioners advising clients who consume energy drinks are working in a different risk category than when discussing filter coffee.

The blood pressure curve is not linear. In people with optimal baseline pressure, consuming one to three cups was slightly associated with a higher hypertension risk. Above three cups, the risk declined again. This sounds paradoxical, but it explains a substantial portion of the apparent contradictions in older studies. For practice, this means: document baseline values rather than issuing blanket warnings.

Preparation method matters. Cafestol is a diterpene that occurs naturally in coffee and measurably raises LDL cholesterol levels. It is present in unfiltered coffee: espresso, French press, Turkish coffee, and moka. Paper filters largely retain cafestol. Practitioners working with clients who have elevated LDL or ApoB should actively ask about preparation method. It is a simple lever that is often overlooked.

The individual factor CYP1A2

Caffeine is metabolised in the liver primarily by the enzyme CYP1A2. Gene variants at the CYP1A2 locus determine whether someone is a fast metaboliser or a slow metaboliser. This difference is clinically relevant.

In slow metabolisers, the half-life of caffeine is significantly extended. Residual caffeine overnight increases. Even moderate consumption can affect sleep architecture, HRV, and blood pressure variability more substantially than in fast metabolisers. Conversely, effects in fast metabolisers are milder and shorter-lived. The same cup of coffee at 4 pm carries fundamentally different consequences for each genotype.

Additional variables further modulate CYP1A2 activity: age, smoking status, pregnancy, hormone therapies, and a range of medications. The enzyme is not a static parameter.

For practice, this means: a caffeine recommendation based on the population average will systematically miss the mark for the individual. Not because the AHA data are wrong, but because they apply to a population — not to the person sitting across from you.

Where SLOW is headed with this

This is precisely where precise health optimisation distinguishes itself from well-intentioned population averages.

Precise means: a coffee recommendation does not end at 400 mg. It begins with the CYP1A2 genotype, is supplemented by ambulatory blood pressure monitoring, sleep architecture, HRV trends, and the lipid profile with a focus on LDL and ApoB. Only then does a recommendation emerge that fits the client — not the study population.

Holistic means: coffee is not an isolated molecule. It sits at the intersection of sleep, stress load, blood glucose regulation, lipid metabolism, and social habit. Preparation method, timing — for slow metabolisers ideally not after 2 pm — additives, and individual context determine whether a cup delivers benefit or side effect. These layers of data cannot be captured with a rule of thumb.

In the SLOW platform, health professionals translate exactly this information into a recommendation that works. The AHA population average becomes the individual threshold. From data to results.

The practitioner remains the expert throughout. SLOW provides the data foundation so that expertise can take effect with greater precision.

How do you educate clients on the topic of coffee — and do you already use CYP1A2 genotyping routinely in practice? Share your approach in the comments.

Source: Marcus GM, Hu FB, van Dam RM, Cornelis MC, Dewland TA, Kang J, Larsson SC, Page RL II, Parekh N. Caffeine and Cardiovascular Disease. A Scientific Statement from the American Heart Association. Circulation (2026). DOI: 10.1161/CIR.0000000000001454