High coffee consumption, brain volume and risk of dementia and stroke

Original Article ↗ Hacker News Discussion ↗

An observational study using UK Biobank data links very high coffee intake (more than six cups per day) to increased odds of dementia compared with moderate consumption of one to two cups, while also finding higher risk among non-drinkers and decaf drinkers. Commenters question causality and highlight potential confounders such as self-medication for ADHD, chronic sleep deprivation, tobacco use, vitamin deficiencies, and the wide variability in what people consider “a cup” of coffee. Many conclude that, at most, the results tentatively favor moderate coffee consumption and underscore the need for better-controlled research and meta-analyses before changing habits.

Interpretation & Causality

  • Several commenters stress the study is observational; it reports associations, not causation.
  • The increased dementia odds for >6 cups/day vs 1–2 cups/day are seen as a signal that needs mechanisms and replication, not behavior-changing proof.
  • Some note the lack of a clear stroke result as a reason to be cautious, calling it potentially a “fishing expedition.”

Sleep, Self‑Medication & Lifestyle Confounders

  • Many suspect heavy coffee use reflects chronic sleep deprivation; poor sleep is repeatedly linked (in the thread) to dementia.
  • Coffee is seen as both cause and effect in a feedback loop: caffeine worsens sleep, bad sleep drives more caffeine.
  • Multiple anecdotes describe coffee (and nicotine) as self‑medication for ADHD or cognitive decline, potentially confounding the coffee–dementia link.
  • High coffee intake is also floated as a proxy for high-stress, low-sleep, mentally demanding lifestyles.

Dose, “Cup” Definition & Preparation Variability

  • Strong concern that “cups” are self‑reported with no size or strength standard; a “cup” might be a small weak instant or a giant strong brew.
  • Commenters note major caffeine differences by brew method, roast, and amount of grounds; some drink huge numbers of very weak cups.
  • This variability is seen as a major limitation for interpreting dose–response.

Caffeine vs. Coffee, Decaf, Tea, Tobacco, Nicotine

  • Several want analyses by total caffeine (including tea and other drinks), not just coffee counts.
  • Discussion highlights that decaf and non‑coffee drinkers also show higher risk than light drinkers in this study, complicating a “caffeine is the toxin” narrative.
  • Coffee–tobacco correlation is mentioned; smoking is proposed as an alternative driver of vascular damage and dementia.
  • Nicotine is framed both as possible self‑medication (especially in ADHD) and as a confounder.

Genetics & Individual Sensitivity

  • Differences in caffeine metabolism genes are discussed as a reason some tolerate large doses while others get anxiety or insomnia from a single cup.
  • This is seen as another unmeasured source of heterogeneity.

Methodology, Other Evidence & Skepticism

  • Some complain that UK Biobank–based diet/health papers generate many associations without explaining “why,” which can unnecessarily alarm people.
  • Another UK Biobank analysis showing coffee linked to increased longevity is cited as seemingly at odds with higher dementia odds at high intakes.
  • A few emphasize waiting for meta‑analyses and expert synthesis; single studies are deemed poor guides for personal decisions.

Personal Experiences & Behavior Changes

  • Numerous anecdotes: from 0–1 highly sensitive cups to pot‑a‑day drinkers who report normal sleep.
  • Several report cutting back or switching to decaf improved sleep and revealed underlying fatigue.
  • Others continue heavy use mainly for taste, while acknowledging possible risks.

Publishing Model & Access

  • Commenters criticize paywalls and odd access models (e.g., 48‑hour paid PDF access) as outdated and obstructive.
  • Workarounds (open repositories, Sci‑Hub) are shared and the cookie/consent UX is mocked.