Results: The two cohorts included 131,821 middle-aged US health professionals, of whom 66 percent were women. Median follow-up at 37 years. Higher coffee consumers tended to be younger, drink more alcohol, smoke more, and consume more total energy. 

Key Result: Health professionals who drank more caffeinated coffee appeared less likely to develop dementia.

• Primary Outcome: Compared with participants in the lowest quartile of caffeinated coffee consumption, those in the highest quartile had a significantly lower risk of incident dementia (hazard ratio (HR) 0.82; 95 percent confidence interval (CI) 0.76 to 0.89). Tea consumption showed a similar association, whereas decaffeinated coffee did not.
  

  Click to enlarge.

• Secondary Outcomes:

Talk Nerdy

1. Residual Confounding and Healthy User Bias: This is the biggest limitation of any observational nutrition study. Even after multivariable adjustment, people who drink more coffee or tea may differ systematically from those who do not in ways that influence cognitive outcomes. These differences may include sleep habits, exercise, education, socioeconomic status, health-seeking behavior, medication use, and overall lifestyle. The authors adjusted for many of these variables, but residual confounding is almost certainly still present. In other words, coffee drinkers may simply differ from non-coffee drinkers in ways that are difficult to fully measure or control.

2. Reverse Causation and Time-Varying Exposure: Early or subclinical cognitive decline could itself change caffeine habits long before a formal dementia diagnosis occurs. People with cognitive impairment may reduce coffee intake due to sleep disturbance, palpitations, gastrointestinal symptoms, or caregiver influence, making coffee appear protective when the disease process alters exposure. This issue is compounded by the fact that coffee and tea consumption likely changed substantially over the 30- to 40-year follow-up period. Even with repeated FFQs, long-term caffeine exposure is difficult to measure accurately over decades.

3. Imperfect Measurement of Exposure and Outcomes: Although repeated validated FFQs are better than a single baseline dietary survey, they still cannot fully capture brew strength, serving size, preparation methods, or changes between caffeinated and decaffeinated products. Similarly, dementia ascertainment relied partly on death records and physician diagnoses rather than standardized in-person cognitive assessments for all participants. These limitations introduce the possibility of exposure misclassification and outcome misclassification, both of which can distort observed associations in longitudinal observational studies.

Bottom Line: Good news for emergency physicians: Your moderate coffee habit may be associated with a lower risk of dementia.

Case Resolution: The emergency physician does not need to panic that their caffeine intake is hurting their brain. It appears that moderate intake of caffeinated coffee or tea is associated with lower dementia risk and slightly better cognitive outcomes, but the observational design cannot establish causation. For emergency physicians who already drink coffee to survive circadian disruption and overnight shifts, this paper suggests their habit is unlikely to harm long-term cognitive health and may even confer some benefit. However, there is no evidence here to recommend escalating caffeine intake beyond moderation, particularly if it worsens anxiety, insomnia, palpitations, reflux, or blood pressure.